serializing non serializable objects in python



~3 min read


460 words

A problem I’ve run into a number of times when working with Python is when I have a datetime object that I want to include in an API call. The first thing I want to do is convert it, along with the rest of my data into a JSON object.

The problem is that when I do that, I get an error: TypeError: Object of type datetime is not JSON serializable!

A quick example:

>>> from datetime import datetime
>>> import json
>>> now =
>>> now
datetime.datetime(2020, 6, 23, 12, 38, 16, 997518)
>>> my_dict = {'now': now}
>>> my_dict
{'now': datetime.datetime(2020, 6, 23, 12, 38, 16, 997518)}
>>> json.dumps(my_dict)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/stephen/.pyenv/versions/3.8.0/lib/python3.8/json/", line 231, in dumps
    return _default_encoder.encode(obj)
  File "/Users/stephen/.pyenv/versions/3.8.0/lib/python3.8/json/", line 199, in encode
    chunks = self.iterencode(o, _one_shot=True)
  File "/Users/stephen/.pyenv/versions/3.8.0/lib/python3.8/json/", line 257, in iterencode
    return _iterencode(o, 0)
  File "/Users/stephen/.pyenv/versions/3.8.0/lib/python3.8/json/", line 179, in default
    raise TypeError(f'Object of type {o.__class__.__name__} '
TypeError: Object of type datetime is not JSON serializable

In this [Stack Overflow](( conversation, a number of solutions are presented, but the more interesting piece is that they all are taking advantage of the same feature of the native JSON encoder in Python: the default argument of the dump and dumps methods.

If specified, default should be a function that gets called for objects that can’t otherwise be serialized. It should return a JSON encodable version of the object or raise a TypeError. If not specified, TypeError is raised.

In their answer, jjmontes makes this very simple and “eats” everything by casting all potentially unserializable elements as a string:

json.dumps(my_dictionary, indent=4, sort_keys=True, default=str)

This works with datetime objects because casting as a string is one of the supported operations for datetime in Python:


Returns a string in the form [D day[s], ][h]H:MM:SS[.UUUUUU], where D is negative for negative t. (5)

A more nuanced approach that only targets datetime objects can be found in jdi’s answer:

import datetime
import json

def default(o):
    if isinstance(o, (, datetime.datetime)):
        return o.isoformat()

return json.dumps(

Or, if you don’t want to lose the error handling altogether, jgbarah’s answer:

from datetime import date, datetime
from json import dumps

def json_serial(obj):
    """JSON serializer for objects not serializable by default json code"""

    if isinstance(obj, (datetime, date)):
        return obj.isoformat()
    raise TypeError ("Type %s not serializable" % type(obj))

print(dumps(, default=json_serial))

The point of all of this, however, is that not all objects are serializable and / or can be cast as strings. It’s possible to use a custom conversion method (by assigning it to the default property) and targeting each type of non serializable object.

This is pretty handy to know!

Hi there and thanks for reading! My name's Stephen. I live in Chicago with my wife, Kate, and dog, Finn. Want more? See about and get in touch!